Skip to main content

03-05-2022 | Rheumatology | News | Article

Data do not support targeted DMARDs for dementia risk reduction

Author: Laura Cowen

medwireNews: Treatment with DMARDs that target cytokine signaling is not associated with a reduced risk for Alzheimer’s disease and related dementia (ADRD) in people with rheumatoid arthritis (RA), US researchers report.

In a previous study – the Drug Repurposing for Effective Alzheimer Medicines (DREAM) initiative – Rishi Desai (Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts) and colleagues identified Janus Kinase (JAK) inhibitors, interleukin (IL)-6 inhibitors, and tumor necrosis factor (TNF) inhibitors as repurposing candidates for reducing ADRD risk based on multiomics phenotyping.

They tested the potential of these candidates relative to the T-cell activation inhibitor abatacept in the current analysis using data for 22,569 propensity score–matched pairs of individuals with RA aged 65 years and older who were treated as Medicare fee-for-service patients between 2007 and 2017.

The cohort included 4224 patients treated with tofacitinib and their matched abatacept controls, 6369 tocilizumab pairs, and 11,976 TNF inhibitor pairs.

To address potential biases including immortal time bias, reverse causation bias, and severe confounding by indication, the researchers used four different analysis schemes. Scheme 1 took an as-treated follow-up approach beginning at cohort entry with follow-up until an ADRD case, treatment discontinuation or switch, death, or the end of the study; scheme 2 used an as-started follow-up approach that included a 6-month lag period and a maximum 3 years of follow-up; scheme 3 incorporated a 6-month symptom to diagnosis period to account for misclassification of ADRD onset; and scheme 4 used an alternative outcome definition that required an ADRD diagnosis code plus at least one symptomatic claim.

The researchers report in JAMA Network Open that across the analysis and treatment groups the incidence of ADRD ranged from 2 to 18 cases per 1000 person–years. The lowest rates were recorded with analysis 4 and the highest rates occurred with analysis 2, but “[r]ates were generally similar within analysis schemes across the 3 treatment cohorts,” Desai et al note.

Furthermore, there were no significant differences in the cumulative incidence of ADRD between each of the three targeted DMARD groups and the abatacept group.

The hazard ratios (HRs) for ADRD were nonsignificant across all comparisons and ranged from 0.50 to 1.29 for tofacitinib versus abatacept, from 0.78 to 1.21 for tocilizumab versus abatacept, and from 0.90 to 1.13 for TNF inhibitors versus abatacept.

Prespecified subgroup analysis by age and sex showed similar results but an analysis by baseline cardiovascular disease (CVD) status suggested that individuals with CVD had a lower risk for ADRD with TNF inhibitors relative to abatacept. The findings were only significant, however, in analyses 2 and 4, with HRs of 0.74 and 0.45, respectively.

The researchers therefore conclude that overall, their “results do not support advancing targeted disease-modifying antirheumatic drugs as disease modifying candidates for ADRD.”

However, they also say that the findings require “a nuanced discussion” as the null findings could be a result of several different factors.

For example, it could be that “DMARDs targeting JAK, IL-6, or TNF, may truly have no causal impact on the risk or trajectory of ADRD” but an alternate explanation could be that these drugs “lower ADRD risk to a similar extent” to that of abatacept, they write.

Nonetheless the investigators believe that using non-use or a nonbiologic DMARD as an alternate comparator would open up the study to “severe bias due to confounding, as the decision to initiate treatment with [targeted] DMARDs in old age is likely influenced by RA activity and frailty, which cannot be fully measured with claims data.”

medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2022 Springer Healthcare Ltd, part of the Springer Nature Group

JAMA Netw Open 2022; 5: e226567

Related topics